INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 calf

-0.72

lac

-0.69

 slee

-0.67

 Lung

-0.65

 teaser

-0.65

tein

-0.64

ellectual

-0.63

 somet

-0.60

 Debor

-0.59

Kau

-0.59

POSITIVE LOGITS

rouse

0.89

iffin

0.76

orthy

0.73

amera

0.71

igham

0.70

este

0.69

 -----

0.69

rir

0.68

pling

0.65

olf

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.